Subtracted X-ray images are derived by subtracting a mask frame (i.e. an image without contrast agent) from one or more fill frames (i.e. an image with contrast agent) to hide (by “subtracting” out) uninteresting anatomical structures (i.e. bones during a vascular study). In order to perform this subtraction the anatomical structures in the mask frame need to be registered (aligned) with the corresponding structures in the fill frames. If a patient being X-rayed moves during the time between the acquisition of a mask frame and acquisition of a fill frame, the registration between the frames may be rendered incorrect and the anatomical structures may not be accurately subtracted out of a displayed composite image.
Known systems support manual and partially automated image shifting enabling a user to select a region of interest (ROI) in an image and automatic calculation of optimal shift values for pixels that fall within this region to align mask and fill frames. In order to store optimal shift values for frames of a multi-frame image sequence, a user is burdened by a need to visit each frame and perform manual tasks (i.e., manual shifting of a mask frame or manual positioning of a ROI). This is a labor and time-intensive operation. A system according to invention principles addresses this problem and related problems.